openfirmware October 2011

openfirmware@openfirmware.info

1 participants
103 discussions

[commit] r2598 - in cpu/arm: . olpc/1.75
by repository service 12 Oct '11

12 Oct '11

Author: wmb Date: Wed Oct 12 23:49:29 2011 New Revision: 2598 URL: http://tracker.coreboot.org/trac/openfirmware/changeset/2598 Log: OLPC XO-1.75 - trac # 11329 - Eliminated, or at least greatly reduced, pops at the beginning and end of the startup sound. The beginning pop was fixed by reordering the CODEC power-up sequence so the DAC is turned on before the output amp. The ending pop was fixed by appending zeros to the upsampling buffer so the filter does not pick up unrelated garbage after the end. Modified: cpu/arm/firfilter.fth cpu/arm/olpc/1.75/alc5631.fth cpu/arm/olpc/1.75/sound.fth Modified: cpu/arm/firfilter.fth ============================================================================== --- cpu/arm/firfilter.fth Wed Oct 12 23:27:50 2011 (r2597) +++ cpu/arm/firfilter.fth Wed Oct 12 23:49:29 2011 (r2598) @@ -21,7 +21,9 @@ \ Taps/phase=16 \ Stride=2 (mono) -code 16tap-upsample ( 'out 'in /in 'weights #phases -- ) +d# 16 constant taps/phase + +code 16tap-upsample ( 'out 'in /in 'weights #phases -- 'out' ) ldmia sp!,{r0,r1,r2,r3} \ r0:'weights r1:/in r2:'in r3:'out tos:#phases mov r4,#15 \ Multiplier scale factor @@ -88,7 +90,7 @@ decs r1,#2 \ Decrement input length by the sample size 0<= until - pop tos,sp + mov tos,r3 c; \ Filter coefficients for 6x upsampling. The following filter was @@ -152,9 +154,19 @@ end [then] +taps/phase 2+ 2* /w* constant /end-buf +/end-buf buffer: end-buf + : upsample6 ( src-adr /src dst-adr -- ) - enable-iwmmx ( src-adr #src-samples dst-adr ) - -rot weights-6phase #phases 16tap-upsample ( ) + enable-iwmmx ( src-adr /src dst-adr ) + >r taps/phase /w* - ( src-adr /src' r: dst-adr ) + r> third third ( src-adr /src dst-adr src-adr /src ) + weights-6phase #phases 16tap-upsample ( src-adr /src dst-adr' ) + -rot + ( dst-adr src-adr' ) + end-buf /end-buf erase ( dst-adr src-adr' ) + end-buf taps/phase /w* move ( dst-adr ) + end-buf taps/phase /w* weights-6phase #phases 16tap-upsample ( dst-adr' ) + drop ; \ LICENSE_BEGIN Modified: cpu/arm/olpc/1.75/alc5631.fth ============================================================================== --- cpu/arm/olpc/1.75/alc5631.fth Wed Oct 12 23:27:50 2011 (r2597) +++ cpu/arm/olpc/1.75/alc5631.fth Wed Oct 12 23:49:29 2011 (r2598) @@ -1,12 +1,6 @@ \ See license at end of file purpose: Driver for Realtek ALC5631Q audio CODEC chip -: adc-on ( -- ) h# 0c00 h# 3a codec-set ; -: adc-off ( -- ) h# 0c00 h# 3a codec-clr ; -: dac-on ( -- ) h# 0300 h# 3a codec-set ; -: dac-off ( -- ) h# 0300 h# 3a codec-clr ; -: adc+dac-on ( -- ) h# 0f00 h# 3a codec-set ; - : set-routing ( -- ) h# c0c0 h# 02 codec-set \ SPKMIXLR -> SPKVOLLR, muted h# c0c0 h# 04 codec-set \ OUTMIXLR -> HPOVOLLR, muted @@ -45,14 +39,6 @@ h# 1010 h# 38 codec! \ Divisors; the values in this register don't seem to make much \ difference unless you set the divisors to very high values. ; -: elided ( -- ) - \ The ADC and DAC will be turned on as needed by adc-on and dac-on, after - \ the BCLK clock from the SoC is on. If you turn on the ADC when BCLK is - \ not clocking, the ADC often doesn't output any data. - b# 1001.0000.1110.0000 h# 3a codec! \ All on except ADC and DAC - b# 1111.1100.0011.1100 h# 3b codec! \ All on except PLL - b# 1111.1100.0000.0000 h# 3e codec! \ AXI and MONO IN off -; : mic-bias-off ( -- ) h# 000c h# 3b codec-clr ; : mic-bias-on ( -- ) h# 000c h# 3b codec-set ; @@ -176,11 +162,13 @@ \ h# 0000 pwr3b! \ Power off PLL ; : open-out-specific ( -- ) + \ Turning on the DAC here at the beginning seems to prevent pops better than doing + \ it a few steps later at the point that Realtek suggested. + h# 0300 h# 3a codec-set \ Power on DACs h# 0060 h# 3a codec-set \ Power on DAC to mixer speakers-on? if h# 1000 h# 3a codec-set then \ Power on ClassD amp speakers-on? if h# c000 h# 3e codec-set then \ Power on SPKL/RVOL headphones-on? if h# 0c00 h# 3e codec-set then \ Power on HPOVOLL/R -\ h# 0300 h# 3a codec-set \ Power on DACL/R - defer until dac-on is called by start-audio-out or out-in h# c000 pwr3b! \ Power on OUTMIXL/R speakers-on? if h# 3000 h# 3b codec-set then \ Power on SPKMIXL/R @@ -192,7 +180,7 @@ open-common open-out-specific ; - + : close-out-specific ( -- ) speakers-on? if mute-speakers then headphones-on? if hp-powerdown-depop then @@ -218,7 +206,7 @@ adc-stereo h# 0c00 h# 3b codec-set \ Power on RECMIXLR h# 0030 h# 3b codec-set \ Power on MIC1/2 boost gain -\ h# 0c00 h# 3a codec-set \ Power on ADCL/R - defer until adc-on is called by audio-in or out-in + h# 0c00 h# 3a codec-set \ Power on ADCL/R ; : open-in ( -- ) \ h# 46f0 h# 44 codec! \ pll: 256000 -> 2048000 ??? why is this different from playback? - 8khz record? Modified: cpu/arm/olpc/1.75/sound.fth ============================================================================== --- cpu/arm/olpc/1.75/sound.fth Wed Oct 12 23:27:50 2011 (r2597) +++ cpu/arm/olpc/1.75/sound.fth Wed Oct 12 23:49:29 2011 (r2598) @@ -345,7 +345,6 @@ out-len if copy-out then \ Prefill the second buffer start-out-ring master-tx - dac-on install-playback-alarm true to playing? ; @@ -390,7 +389,6 @@ make-in-ring ( actual ) start-in-ring ( actual ) master-rx ( actual ) - adc-on ( actual ) begin in-len while ( actual ) wait-in ( actual ) copy-in ( actual ) @@ -436,8 +434,6 @@ master-rx ( ) \ Now the clock is on slave-tx ( ) - adc+dac-on ( ) - true to playing? begin in-len playing? or while ( ) @@ -453,11 +449,8 @@ reset-rx reset-tx - dac-off adc-off ( ) - - mono? if collapse-in then ( ) - close-out-in + mono? if collapse-in then ( ) ; 0 [if] \ Interactive test words for out-in

1 0

[commit] r2597 - cpu/x86/pc/olpc/via
by repository service 12 Oct '11

12 Oct '11

Author: quozl Date: Wed Oct 12 23:27:50 2011 New Revision: 2597 URL: http://tracker.coreboot.org/trac/openfirmware/changeset/2597 Log: Q3B20 (note that this version was not svn head at this point, but rather svn 2318 with 2589) Modified: cpu/x86/pc/olpc/via/ec-version.fth cpu/x86/pc/olpc/via/fw-version.fth Modified: cpu/x86/pc/olpc/via/ec-version.fth ============================================================================== --- cpu/x86/pc/olpc/via/ec-version.fth Wed Oct 12 20:49:37 2011 (r2596) +++ cpu/x86/pc/olpc/via/ec-version.fth Wed Oct 12 23:27:50 2011 (r2597) @@ -1,5 +1,5 @@ \ The EC microcode -macro: EC_VERSION 2_2_8 +macro: EC_VERSION 2_2_9 \ Alternate command for getting EC microcode, for testing new versions. \ Temporarily uncomment the line and modify the path as necessary Modified: cpu/x86/pc/olpc/via/fw-version.fth ============================================================================== --- cpu/x86/pc/olpc/via/fw-version.fth Wed Oct 12 20:49:37 2011 (r2596) +++ cpu/x86/pc/olpc/via/fw-version.fth Wed Oct 12 23:27:50 2011 (r2597) @@ -1,3 +1,3 @@ \ The overall firmware revision macro: FW_MAJOR B -macro: FW_MINOR 19 +macro: FW_MINOR 20

1 0

[commit] r2596 - cpu/arm/olpc/1.75
by repository service 12 Oct '11

12 Oct '11

Author: wmb Date: Wed Oct 12 20:49:37 2011 New Revision: 2596 URL: http://tracker.coreboot.org/trac/openfirmware/changeset/2596 Log: OLPC XO-1.75 - fixed ATAG botch introduced by last revision of fw.bth - prevented booting recent Linux builds. Modified: cpu/arm/olpc/1.75/fw.bth Modified: cpu/arm/olpc/1.75/fw.bth ============================================================================== --- cpu/arm/olpc/1.75/fw.bth Wed Oct 12 03:40:18 2011 (r2595) +++ cpu/arm/olpc/1.75/fw.bth Wed Oct 12 20:49:37 2011 (r2596) @@ -71,13 +71,13 @@ cif-handler tag-l, \ Client interface handler callback address \ Each of these groups is a struct map_desc as defined in arch/arm/include/asm/mach/ - extra-mem-va tag-l, \ VA of OFW memory - dup >physical pageshift rshift tag-l, \ Page frame number of OFW memory + extra-mem-va dup tag-l, \ VA of OFW memory + >physical pageshift rshift tag-l, \ Page frame number of OFW memory fw-mem-va /fw-mem + extra-mem-va - tag-l, \ Size of OFW memory MT_MEMORY tag-l, \ Mapping type of OFW memory - fb-mem-va tag-l, \ VA of OFW Frame Buffer - dup >physical pageshift rshift tag-l, \ PA of OFW Frame Buffer + fb-mem-va dup tag-l, \ VA of OFW Frame Buffer + >physical pageshift rshift tag-l, \ PA of OFW Frame Buffer /fb-mem tag-l, \ Size of OFW memory MT_DEVICE_WC tag-l, \ Mapping type of OFW frame buffer ;

1 0

[commit] r2595 - cpu/arm/olpc/1.75
by repository service 12 Oct '11

12 Oct '11

Author: wmb Date: Wed Oct 12 03:40:18 2011 New Revision: 2595 URL: http://tracker.coreboot.org/trac/openfirmware/changeset/2595 Log: OLPC XO-1.75 - Minor cleanups in accelerometer driver. Modified: cpu/arm/olpc/1.75/accelerometer.fth Modified: cpu/arm/olpc/1.75/accelerometer.fth ============================================================================== --- cpu/arm/olpc/1.75/accelerometer.fth Wed Oct 12 02:48:03 2011 (r2594) +++ cpu/arm/olpc/1.75/accelerometer.fth Wed Oct 12 03:40:18 2011 (r2595) @@ -8,7 +8,6 @@ \ This is for the stand-alone accelerometer chip LIS33DETR \ We could call this just once in open if we had a TWSI parent node -: set-address ( -- ) h# 3a 6 set-twsi-target ; : acc-reg@ ( reg# -- b ) 1 1 " smbus-out-in" $call-parent ; : acc-reg! ( b reg# -- ) 2 0 " smbus-out-in" $call-parent ; : ctl1! ( b -- ) h# 20 acc-reg! ; @@ -46,7 +45,7 @@ : delay ( -- ) d# 30 ms ; : open ( -- flag ) - my-address my-space " set-address" $call-parent + my-unit " set-address" $call-parent d# 25,000 " set-bus-speed" $call-parent ['] accelerometer-on catch 0= ;

1 0

[commit] r2594 - cpu/x86/pc/olpc
by repository service 12 Oct '11

12 Oct '11

Author: quozl Date: Wed Oct 12 02:48:03 2011 New Revision: 2594 URL: http://tracker.coreboot.org/trac/openfirmware/changeset/2594 Log: Q2E47 (note that this version was not svn head at this point, but rather svn 2393 with svn 2585 and 2589) Modified: cpu/x86/pc/olpc/versions.fth Modified: cpu/x86/pc/olpc/versions.fth ============================================================================== --- cpu/x86/pc/olpc/versions.fth Wed Oct 12 02:30:09 2011 (r2593) +++ cpu/x86/pc/olpc/versions.fth Wed Oct 12 02:48:03 2011 (r2594) @@ -2,7 +2,7 @@ \ The overall firmware revision macro: FW_MAJOR E -macro: FW_MINOR 46 +macro: FW_MINOR 47 \ The EC microcode macro: EC_VERSION e35

1 0

[commit] r2593 - in cpu/arm: mmp2 olpc olpc/1.75
by repository service 12 Oct '11

12 Oct '11

Author: wmb Date: Wed Oct 12 02:30:09 2011 New Revision: 2593 URL: http://tracker.coreboot.org/trac/openfirmware/changeset/2593 Log: OLPC XO-1.75 - Eliminated hardcoded DRAM physical addresses. In conjunction with a change to CForth, this should permit automatic support of different memory sizes. Modified: cpu/arm/mmp2/mmuon.fth cpu/arm/mmp2/rootnode.fth cpu/arm/olpc/1.75/addrs.fth cpu/arm/olpc/1.75/devices.fth cpu/arm/olpc/1.75/fw.bth cpu/arm/olpc/1.75/lcd.fth cpu/arm/olpc/1.75/prefw.bth cpu/arm/olpc/1.75/probemem.fth cpu/arm/olpc/initmmu.fth cpu/arm/olpc/resetvec.bth Modified: cpu/arm/mmp2/mmuon.fth ============================================================================== --- cpu/arm/mmp2/mmuon.fth Wed Oct 12 00:29:49 2011 (r2592) +++ cpu/arm/mmp2/mmuon.fth Wed Oct 12 02:30:09 2011 (r2593) @@ -9,9 +9,11 @@ mov tos,tos,lsl #14 c; +: (page-table-va) fw-mem-va page-table-offset + ; +' (page-table-va) to page-table-va : map-section ( pa+mode va -- ) - d# 18 rshift page-table@ + tuck l! clean-d$-entry + d# 18 rshift page-table-va + tuck l! clean-d$-entry ; : map-sections ( pa mode va size -- ) 2>r + 2 or 2r> ( pa+mode va size ) @@ -22,6 +24,7 @@ drop ; +[ifdef] notdef : ofw-sections ( -- ) h# 0000.0000 h# c0e over dma-mem-pa map-sections \ Cache and write bufferable dma-mem-pa h# c02 over /dma-mem map-sections \ Non-cacheable DMA space @@ -34,14 +37,6 @@ h# e000.0000 h# c02 over /section map-sections \ Audio SRAM - no caching or buffering ; -: setup-sections - page-table-pa page-table! - page-table-pa /page-table erase - - ofw-sections -; -\ Do we need to map SRAM and DDRC ? - code start-mmu set r2, 0xFFFFFFFF \ Set domains for Manager access mcr p15,0,r2,3,0,0 \ Update register 3 in CP15 @@ -67,10 +62,21 @@ sub pc, pc, #4 c; +: setup-sections + ['] page-table@ to page-table-va + page-table-pa dup page-table! /page-table erase + + ofw-sections + ['] (page-table-va) to page-table-va + start-mmu +; +[then] + : go-fast control@ 1 and if exit then \ Don't do this if MMU is already on +[ifdef] notdef setup-sections - start-mmu +[then] dcache-on icache-on \ l2cache-on \ Leave off for now, to avoid potential problems with Linux Modified: cpu/arm/mmp2/rootnode.fth ============================================================================== --- cpu/arm/mmp2/rootnode.fth Wed Oct 12 00:29:49 2011 (r2592) +++ cpu/arm/mmp2/rootnode.fth Wed Oct 12 02:30:09 2011 (r2593) @@ -26,14 +26,15 @@ : map-in ( phys size -- virt ) drop ( phys ) - dup fb-mem-pa = if drop fb-mem-va exit then ( phys ) + \ The display driver uses fb-mem-va directly instead of calling map-in + \ dup fb-mem-va >physical = if drop fb-mem-va exit then ( phys ) io-pa - io-va + ; : map-out ( virtual size -- ) 2drop ; -: dma-range ( -- start end ) dma-mem-pa dup /dma-mem + ; +: dma-range ( -- start end ) dma-mem-va >physical dup /dma-mem + ; h# 0 constant dma-map-mode \ XXX what should this be? @@ -92,7 +93,10 @@ mmu-map ( ) r> ( virt ) [else] - nip + nip ( phys ) +[ifdef] dma-mem-va + dma-mem-va >physical - dma-mem-va + ( virt ) +[then] [then] ; : dma-free ( virt size -- ) @@ -102,21 +106,30 @@ ( virt size phys ) -rot tuck ( phys size virt size ) 2dup mmu-unmap mmu-release ( phys size ) +[else] +[ifdef] >physical + swap >physical swap ( virt ) +[then] [then] mem-release ( ) ; : dma-map-in ( virt size cacheable -- devaddr ) - drop 2dup flush-d$-range drop ( virt ) + drop ( virt size ) + 2dup flush-d$-range ( virt size ) + drop ( virt ) +[ifdef] >physical + >physical ( phys ) +[then] ; : dma-map-out ( virt devaddr size -- ) nip flush-d$-range ; \ : dma-sync ( virt devaddr size -- ) nip flush-d$-range ; \ : dma-push ( virt devaddr size -- ) nip flush-d$-range ; \ : dma-pull ( virt devaddr size -- ) nip flush-d$-range ; -: dma-sync ( virt devaddr size -- ) 3drop ; +: dma-sync ( virt devaddr size -- ) 3drop d# 30 us ; : dma-push ( virt devaddr size -- ) 3drop ; -: dma-pull ( virt devaddr size -- ) 3drop ; +: dma-pull ( virt devaddr size -- ) 3drop d# 30 us ; finish-device Modified: cpu/arm/olpc/1.75/addrs.fth ============================================================================== --- cpu/arm/olpc/1.75/addrs.fth Wed Oct 12 00:29:49 2011 (r2592) +++ cpu/arm/olpc/1.75/addrs.fth Wed Oct 12 02:30:09 2011 (r2593) @@ -2,24 +2,12 @@ fload ${BP}/cpu/arm/mmuparams.fth -h# 2000.0000 constant /ram-total \ Total size of memory - h# 0040.0000 constant /fb-mem \ The screen uses a little more than 3 MiB at 1200x900x24 -/ram-total /fb-mem - constant fb-mem-pa \ e.g. h# 1fc0.0000 - -fb-mem-pa constant /available-mem - -: (memory?) ( phys -- flag ) /ram-total u< ; \ OFW implementation choices -h# 0020.0000 constant /fw-mem -fb-mem-pa /fw-mem - constant fw-mem-pa \ e.g. h# 1fa0.0000 - -h# 0020.0000 constant /extra-mem -fw-mem-pa /extra-mem - constant extra-mem-pa \ e.g. h# 1f80.0000 - -h# 0080.0000 constant /dma-mem -extra-mem-pa /dma-mem - constant dma-mem-pa \ e.g. h# 1f00.0000 +h# 0020.0000 constant /fw-mem +h# 0020.0000 constant /extra-mem +h# 0080.0000 constant /dma-mem h# fd00.0000 constant dma-mem-va h# fd80.0000 constant extra-mem-va @@ -39,7 +27,7 @@ /fw-mem /page-table - constant page-table-offset page-table-offset constant stack-offset \ Stack is below this -fw-mem-pa page-table-offset + constant page-table-pa +\ fw-mem-pa page-table-offset + constant page-table-pa \ h# 0110.0000 constant def-load-base h# 0800.0000 constant def-load-base Modified: cpu/arm/olpc/1.75/devices.fth ============================================================================== --- cpu/arm/olpc/1.75/devices.fth Wed Oct 12 00:29:49 2011 (r2592) +++ cpu/arm/olpc/1.75/devices.fth Wed Oct 12 02:30:09 2011 (r2593) @@ -242,14 +242,17 @@ init-lcd ; : map-frame-buffer ( -- ) - fb-mem-pa /fb-mem " map-in" $call-parent to frame-buffer-adr + \ We use fb-mem-va directly instead of calling map-in on the physical address + \ because the physical address changes with the total memory size. The early + \ assembly language startup code establishes the mapping. + fb-mem-va to frame-buffer-adr ; " display" device-type " ISO8859-1" encode-string " character-set" property 0 0 encode-bytes " iso6429-1983-colors" property \ Used as temporary storage for images by $get-image - : graphmem ( -- adr ) dimensions * pixel* fb-mem-pa + ; + : graphmem ( -- adr ) dimensions * pixel* fb-mem-va + ; : display-install ( -- ) map-frame-buffer @@ -287,8 +290,6 @@ fload ${BP}/dev/olpc/kb3700/spicmd.fth fload ${BP}/cpu/arm/olpc/spcmd.fth -devalias keyboard /ec-spi/keyboard - : wlan-reset ( -- ) d# 58 gpio-clr d# 20 ms d# 58 gpio-set ; \ Create the alias unless it already exists Modified: cpu/arm/olpc/1.75/fw.bth ============================================================================== --- cpu/arm/olpc/1.75/fw.bth Wed Oct 12 00:29:49 2011 (r2592) +++ cpu/arm/olpc/1.75/fw.bth Wed Oct 12 02:30:09 2011 (r2593) @@ -72,12 +72,12 @@ \ Each of these groups is a struct map_desc as defined in arch/arm/include/asm/mach/ extra-mem-va tag-l, \ VA of OFW memory - extra-mem-pa pageshift rshift tag-l, \ Page frame number of OFW memory + dup >physical pageshift rshift tag-l, \ Page frame number of OFW memory fw-mem-va /fw-mem + extra-mem-va - tag-l, \ Size of OFW memory MT_MEMORY tag-l, \ Mapping type of OFW memory fb-mem-va tag-l, \ VA of OFW Frame Buffer - fb-mem-pa pageshift rshift tag-l, \ PA of OFW Frame Buffer + dup >physical pageshift rshift tag-l, \ PA of OFW Frame Buffer /fb-mem tag-l, \ Size of OFW memory MT_DEVICE_WC tag-l, \ Mapping type of OFW frame buffer ; Modified: cpu/arm/olpc/1.75/lcd.fth ============================================================================== --- cpu/arm/olpc/1.75/lcd.fth Wed Oct 12 00:29:49 2011 (r2592) +++ cpu/arm/olpc/1.75/lcd.fth Wed Oct 12 02:30:09 2011 (r2593) @@ -16,7 +16,7 @@ lcd-clocks-on 0 h# 190 lcd! \ Disable LCD DMA controller - fb-mem-pa h# f4 lcd! \ Frame buffer area 0 + fb-mem-va >physical h# f4 lcd! \ Frame buffer area 0 0 h# f8 lcd! \ Frame buffer area 1 hdisp bytes/pixel * h# fc lcd! \ Pitch in bytes Modified: cpu/arm/olpc/1.75/prefw.bth ============================================================================== --- cpu/arm/olpc/1.75/prefw.bth Wed Oct 12 00:29:49 2011 (r2592) +++ cpu/arm/olpc/1.75/prefw.bth Wed Oct 12 02:30:09 2011 (r2593) @@ -12,7 +12,6 @@ ' noop is include-hook fload ${BP}/cpu/arm/olpc/1.75/config.fth -' (memory?) to memory? : headerless ; : headers ; : headerless0 ; @@ -33,37 +32,38 @@ fload ${BP}/ofw/core/memlist.fth \ Resource list common routines fload ${BP}/ofw/core/showlist.fth \ Linked list display tool -fload ${BP}/cpu/arm/mmp2/rootnode.fth \ Root node mapping - physical mode -dev / - " olpc,XO-1.75" model - " Marvell,Armada 610" encode-string " architecture" property -\ The clock frequency of the root bus may be irrelevant, since the bus is internal to the SOC -\ d# 1,000,000,000 " clock-frequency" integer-property -device-end - -: (cpu-arch ( -- adr len ) - " architecture" ['] root-node get-package-property drop - get-encoded-string -; -' (cpu-arch to cpu-arch - \ Memory management services [ifdef] virtual-mode fload ${BP}/ofw/core/clntmem1.fth \ client services for memory [else] fload ${BP}/ofw/core/clntphy1.fth \ client services for memory -defer section-table +defer page-table-va : >physical ( va -- pa ) dup d# 20 rshift ( va section-index ) - section-table swap la+ l@ ( va pte ) + page-table-va swap la+ l@ ( va pte ) h# fffff invert and ( va pa-base ) swap h# fffff and or ( pa ) ; [then] +fload ${BP}/cpu/arm/mmp2/rootnode.fth \ Root node mapping - physical mode + fload ${BP}/ofw/core/allocph1.fth \ S Physical memory allocator fload ${BP}/ofw/core/availpm.fth \ Available memory list +dev / + " olpc,XO-1.75" model + " Marvell,Armada 610" encode-string " architecture" property +\ The clock frequency of the root bus may be irrelevant, since the bus is internal to the SOC +\ d# 1,000,000,000 " clock-frequency" integer-property +device-end + +: (cpu-arch ( -- adr len ) + " architecture" ['] root-node get-package-property drop + get-encoded-string +; +' (cpu-arch to cpu-arch + fload ${BP}/cpu/arm/olpc/1.75/probemem.fth \ Memory probing stand-init: Probing memory @@ -79,7 +79,6 @@ fload ${BP}/arch/arm/loadarea.fth \ Allocate and map program load area [else] fload ${BP}/cpu/arm/mmp2/mmuon.fth -' page-table@ to section-table [then] \ XXX should be elsewhere Modified: cpu/arm/olpc/1.75/probemem.fth ============================================================================== --- cpu/arm/olpc/1.75/probemem.fth Wed Oct 12 00:29:49 2011 (r2592) +++ cpu/arm/olpc/1.75/probemem.fth Wed Oct 12 02:30:09 2011 (r2593) @@ -3,6 +3,9 @@ " /memory" find-device +: (memory?) ( phys -- flag ) fb-mem-va >physical u< ; +' (memory?) to memory? + headerless h# ffff.ffff value low @@ -16,15 +19,16 @@ headers : probe ( -- ) - 0 /available-mem log&release + 0 fb-mem-va >physical /fb-mem + log&release 0 0 encode-bytes ( adr 0 ) physavail ['] make-phys-memlist find-node ( adr len prev 0 ) 2drop " reg" property - \ Claim the memory used by OFW - fw-mem-pa /fw-mem 0 claim drop - extra-mem-pa /extra-mem 0 claim drop + \ Claim the memory already in use + fb-mem-va >physical /fb-mem 0 claim drop + fw-mem-va >physical /fw-mem 0 claim drop + extra-mem-va >physical /extra-mem 0 claim drop 0 pagesize 0 claim drop \ Vector table ; Modified: cpu/arm/olpc/initmmu.fth ============================================================================== --- cpu/arm/olpc/initmmu.fth Wed Oct 12 00:29:49 2011 (r2592) +++ cpu/arm/olpc/initmmu.fth Wed Oct 12 02:30:09 2011 (r2593) @@ -204,12 +204,61 @@ mov pc, lr end-code +\ Map sections within the given address range, using +\ the given protection/cacheability mode. pt-adr is the page table base address. +label allocate-and-map-sections ( r0: pt-adr, r1: padr-top, r2: len, r3: mode r4: vadr -- r1: padr-bot ) + inc r4, r2 \ vadr-top + add r1, r1, r3 \ PA+mode + begin + dec r1, #0x100000 + dec r4, #0x100000 + + str r1, [r0, r4, lsr #18] + + decs r2, #0x100000 + 0<= until + + mov r1,r1,lsr #20 \ Clear out mode bits in padr + mov r1,r1,lsl #20 + + mov pc, lr +end-code + +\ This assumes that there are no holes and that unused MMAP registers have CS_VALID=0 +label dramsize ( -- r0: size-in-bytes ) + mov r0,0 + mov r1,0xd0000000 \ Memory controller base address + + ldr r2,[r1,#0x100] \ MMAP0 register + ands r3,r2,#1 \ Test CS_VALID + movne r3,#0x10000 \ Scale factor for memory size + movne r2,r2,lsl #12 \ Clear high bits above AREA_LENGTH field + movne r2,r2,lsr #28 \ Move AREA_LENGTH to LSB + addne r0,r0,r3,lsl r2 \ Compute bank size and add to accumulator + + ldr r2,[r1,#0x110] \ MMAP1 register + ands r3,r2,#1 \ Test CS_VALID + movne r3,#0x10000 \ Scale factor for memory size + movne r2,r2,lsl #12 \ Clear high bits above AREA_LENGTH field + movne r2,r2,lsr #28 \ Move AREA_LENGTH to LSB + addne r0,r0,r3,lsl r2 \ Compute bank size and add to accumulator + + mov pc,lr +end-code + \ Initial the section table, setting up mappings for the platform-specific \ address ranges that the firmware uses. \ Destroys: r0-r4 -label init-map ( r0: section-table -- ) +label init-map ( -- ) mov r10,lr + bl `dramsize` \ r0: total-memory-size + mov r1,r0 \ r1: allocation pointer starts at top of DRAM + + \ Locate the page table at the top of the firmware memory, just below the frame buffer + dec r0,`/fb-mem #` \ Size of frame buffer + dec r0,`/page-table #` \ r0: page-table-pa + mcr p15,0,r0,cr2,cr0,0 \ Set table base address \ Clear the entire section table for starters @@ -219,32 +268,33 @@ subs r2, r2, #1 \ Decrement section number str r3, [r0, r2, lsl #2] \ Invalidate section entry 0= until + \ r1: top of DRAM + set r2,`/fb-mem #` \ Size of frame buffer + set r3,#0xc06 \ Write bufferable + set r4,`fb-mem-va #` \ Virtual address + bl `allocate-and-map-sections \ r1: bottom PA of frame buffer - mov r1,0 \ Address of low memory - set r2,`dma-mem-pa #` \ Size of low memory - up to dma-base - set r3,#0xc0e \ Cache and write bufferable - bl `map-sections-v=p` - - set r1,`dma-mem-pa #` \ Address of DMA area - set r2,`/dma-mem #` \ Size of DMA area - set r3,#0xc02 \ No caching or write buffering - bl `map-sections-v=p` + set r2,`/fw-mem #` \ Size of firmware region + set r3,#0xc0e \ Write bufferable + set r4,`fw-mem-va #` \ Virtual address + bl `allocate-and-map-sections` \ r1: bottom PA of firmware memory - set r1,`extra-mem-pa #` \ Address of additional allocatable memory set r2,`/extra-mem #` \ Size of additional allocatable memory set r3,#0xc0e \ Write bufferable - bl `map-sections-v=p` + set r4,`extra-mem-va #' \ Virtual address + bl `allocate-and-map-sections` \ r1: bottom PA of extra memory - set r1,`fw-mem-pa #` \ Address of Firmware region - set r2,`/fw-mem #` \ Size of firmware region - set r3,#0xc0e \ Write bufferable - bl `map-sections-v=p` + set r2,`/dma-mem #` \ Size of DMA area + set r3,#0xc02 \ No caching or write buffering + set r4,`dma-mem-va #` \ Virtual address + bl `allocate-and-map-sections` \ r1: bottom PA of DMA memory - set r1,`fb-mem-pa #` \ Address - Frame buffer - set r2,`/fb-mem #` \ Size of frame buffer - set r3,#0xc06 \ Write bufferable - bl `map-sections-v=p` + mov r2,r1 \ Size of low memory + set r3,#0xc0e \ Cache and write bufferable + mov r4,#0 \ Virtual address + bl `allocate-and-map-sections` \ r1: 0 + \ Now we have mapped all of DRAM set r1,`sram-pa #` \ Address of SRAM set r2,`/sram #` \ Size of SRAM set r3,#0xc02 \ No caching or write buffering @@ -260,40 +310,12 @@ set r3,#0xc02 \ No caching or write buffering bl `map-sections-v=p` - set r1,`dma-mem-pa #` \ Address of DMA area - set r2,`/dma-mem #` \ Size of DMA area - set r3,#0xc02 \ No caching or write buffering - set r4,`dma-mem-va #` \ Virtual address - bl `map-sections` - - set r1,`fb-mem-pa #` \ Address - Frame buffer - set r2,`/fb-mem #` \ Size of frame buffer - set r3,#0xc06 \ Write bufferable - set r4,`fb-mem-va #` \ Virtual address - bl `map-sections - - set r1,`extra-mem-pa #` \ Address of additional allocatable memory - set r2,`/extra-mem #` \ Size of additional allocatable memory - set r3,#0xc0e \ Write bufferable - set r4,`extra-mem-va #' \ Virtual address - bl `map-sections` - - set r1,`fw-mem-pa #` \ Address of Firmware region - set r2,`/fw-mem #` \ Size of firmware region - set r3,#0xc0e \ Write bufferable - set r4,`fw-mem-va #` \ Virtual address - bl `map-sections` - set r1,`io-pa #` \ Address of I/O set r2,`/io #` \ Size of I/O region set r3,#0xc02 \ No caching or write buffering set r4,`io-va #` \ Virtual address bl `map-sections` -\ The cache is not on yet -\ set r1,#0x4000 \ Size of section table -\ bl `clean-dcache-range` - mov pc, r10 end-code Modified: cpu/arm/olpc/resetvec.bth ============================================================================== --- cpu/arm/olpc/resetvec.bth Wed Oct 12 00:29:49 2011 (r2592) +++ cpu/arm/olpc/resetvec.bth Wed Oct 12 02:30:09 2011 (r2593) @@ -152,7 +152,6 @@ \ Setup the page (section) table and turn on the MMU and caches \ set r0,`page-table-pa #` - set r0,`fw-mem-pa page-table-offset + #` bl `init-map` \ Setup the initial virtual address map bl `enable-mmu` \ Turn on the MMU bl `caches-on` \ Turn on the caches

1 0

[commit] r2592 - in cpu/x86/pc/olpc: . via
by repository service 12 Oct '11

12 Oct '11

Author: wmb Date: Wed Oct 12 00:29:49 2011 New Revision: 2592 URL: http://tracker.coreboot.org/trac/openfirmware/changeset/2592 Log: OLPC - changed the IP address for some developer-convenience functions. Only affects Mitch. Modified: cpu/x86/pc/olpc/fw.bth cpu/x86/pc/olpc/via/fw.bth Modified: cpu/x86/pc/olpc/fw.bth ============================================================================== --- cpu/x86/pc/olpc/fw.bth Wed Oct 12 00:28:40 2011 (r2591) +++ cpu/x86/pc/olpc/fw.bth Wed Oct 12 00:29:49 2011 (r2592) @@ -644,11 +644,7 @@ quit ; -: newrom - " flash! http:\\10.20.0.14\new.rom" eval -\ " wifi media lab 802.11" eval -\ " flash http:\\18.85.46.172\new.rom" eval -; +: newrom " flash! http:\\192.168.200.200\new.rom" eval ; : urom " flash! u:\new.rom" eval ; \ Fancy battery charge logger. Modified: cpu/x86/pc/olpc/via/fw.bth ============================================================================== --- cpu/x86/pc/olpc/via/fw.bth Wed Oct 12 00:28:40 2011 (r2591) +++ cpu/x86/pc/olpc/via/fw.bth Wed Oct 12 00:29:49 2011 (r2592) @@ -482,8 +482,9 @@ end-support-package devalias smb tcp//cifs devalias cifs tcp//cifs -: op " select smb:\\test:testxxx@10.20.0.14\XTest\hello.txt" eval ; -: dsmb " dir smb:\\test:testxxx@10.20.0.14\XTest\" eval ; +\ Examples: +\ select smb:\\test:testxxx@10.20.0.14\XTest\hello.txt +\ dir smb:\\test:testxxx@10.20.0.14\XTest\ fload ${BP}/cpu/x86/pc/olpc/via/memtest.fth @@ -733,9 +734,7 @@ : newrom enable-serial - " flash! http:\\10.20.0.14\new.rom" eval -\ " wifi media lab 802.11" eval -\ " flash http:\\18.85.46.172\new.rom" eval + " flash! http:\\192.168.200.200\new.rom" eval ; : urom enable-serial " flash! u:\new.rom" eval ; : qrom enable-serial " essid qz" eval " flash http:\\olpc.lan\qrom.rom" eval ;

1 0

[commit] r2591 - in dev/usb2: . hcd hcd/ehci hcd/ohci hcd/uhci
by repository service 12 Oct '11

12 Oct '11

Author: wmb Date: Wed Oct 12 00:28:40 2011 New Revision: 2591 URL: http://tracker.coreboot.org/trac/openfirmware/changeset/2591 Log: USB - Use interrupt-on-completion for bulk and control transfers, thus eliminating or at least greatly reducing problems related to descriptor polling causing DMA failures. Also use dma-push and dma-pull instead of dma-sync, for more precise control of coherency. Modified: dev/usb2/align.fth dev/usb2/hcd/control.fth dev/usb2/hcd/ehci/bulk.fth dev/usb2/hcd/ehci/control.fth dev/usb2/hcd/ehci/ehci.fth dev/usb2/hcd/ehci/intr.fth dev/usb2/hcd/ehci/iso.fth dev/usb2/hcd/ehci/qhtd.fth dev/usb2/hcd/ohci/bulk.fth dev/usb2/hcd/ohci/control.fth dev/usb2/hcd/ohci/edtd.fth dev/usb2/hcd/ohci/intr.fth dev/usb2/hcd/uhci/bulk.fth dev/usb2/hcd/uhci/control.fth dev/usb2/hcd/uhci/intr.fth dev/usb2/hcd/uhci/qhtd.fth Modified: dev/usb2/align.fth ============================================================================== --- dev/usb2/align.fth Tue Oct 11 23:59:39 2011 (r2590) +++ dev/usb2/align.fth Wed Oct 12 00:28:40 2011 (r2591) @@ -11,7 +11,8 @@ : round-up ( n align -- n' ) 1- tuck + swap invert and ; external -: dma-sync ( virt phys size -- ) " dma-sync" $call-parent ; +: dma-push ( virt phys size -- ) " dma-push" $call-parent ; +: dma-pull ( virt phys size -- ) " dma-pull" $call-parent ; : dma-alloc ( size -- virt ) " dma-alloc" $call-parent ; : dma-free ( virt size -- ) " dma-free" $call-parent ; : dma-map-in ( virt size cache? -- phys ) " dma-map-in" $call-parent ; Modified: dev/usb2/hcd/control.fth ============================================================================== --- dev/usb2/hcd/control.fth Tue Oct 11 23:59:39 2011 (r2590) +++ dev/usb2/hcd/control.fth Wed Oct 12 00:28:40 2011 (r2591) @@ -14,7 +14,7 @@ tuck >dr-value le-w! ( len idx vpcbp ) tuck >dr-index le-w! ( len vpcbp ) >dr-len le-w! ( ) - setup-buf setup-buf-phys /dr dma-sync ( ) + setup-buf setup-buf-phys /dr dma-push ( ) ; external Modified: dev/usb2/hcd/ehci/bulk.fth ============================================================================== --- dev/usb2/hcd/ehci/bulk.fth Tue Oct 11 23:59:39 2011 (r2590) +++ dev/usb2/hcd/ehci/bulk.fth Wed Oct 12 00:28:40 2011 (r2591) @@ -225,6 +225,7 @@ dup le-l@ ( in? token-adr token-val r: qtd qh ) rot if bulk-in-data@ else bulk-out-data@ then ( token-adr token-val toggle r: qtd qh ) or ( token-adr token-val' r: qtd qh ) + TD_IOC or ( token-adr token-val' r: qtd qh ) swap le-l! ( r: qtd qh ) r> r> ( qh qtd ) @@ -257,7 +258,7 @@ over >hcqtd-token ( qtd len token-adr ) tuck 2+ le-w! ( qtd token-adr ) TD_C_ERR3 TD_PID_OUT or TD_STAT_PING or TD_STAT_ACTIVE or swap le-w! ( qtd ) - sync-qtd + push-qtd ; : wait-out ( qtd -- error? ) @@ -310,12 +311,12 @@ : bulk-in-ready? ( -- false | error true | buf actual 0 true ) clear-usb-error bulk-in-qtd >r - r@ sync-qtd + r@ pull-qtd r@ qtd-done? if ( ) r@ bulk-in-qh qtd-error? ?dup 0= if ( ) r@ >qtd-buf l@ ( buf actual ) r@ qtd-get-actual ( buf actual ) - 2dup r@ >qtd-pbuf l@ swap dma-sync ( buf actual ) + 2dup r@ >qtd-pbuf l@ swap dma-pull ( buf actual ) 0 ( buf actual 0 ) then ( error | buf actual 0 ) true ( ... ) @@ -352,7 +353,7 @@ \ Recycle the first qtd last so the transaction is atomic WRT the HC drop dup recycle-one-qtd ( qtd0 ) - sync-qtds + push-qtds ; \ Fixup the host-controller-writable fields in the chain of qTDs - @@ -451,6 +452,12 @@ : recycle-buffer restart-bulk-in ; +: start-bulk-transaction ( pid -- ) + my-bulk-qtd fill-bulk-io-qtds ( ) + my-bulk-qh pt-bulk fill-qh ( ) + my-bulk-qh interrupt-on-last-td ( ) + my-bulk-qh insert-qh ( ) +; : bulk-in ( buf len pipe -- actual usberr ) debug? if ." bulk-in" cr then dup to bulk-in-pipe @@ -459,22 +466,14 @@ bulk-in-timeout my-bulk-qh >qh-timeout l! \ IN qTDs - TD_PID_IN my-bulk-qtd fill-bulk-io-qtds - - \ Start bulk in transaction - my-bulk-qh pt-bulk fill-qh - my-bulk-qh insert-qh + TD_PID_IN start-bulk-transaction ( ) \ Process results - my-bulk-qh done? if ( ) - 0 ( actual ) \ System error, timeout + my-bulk-qh done-error? if ( ) \ System error, timeout, or USB error + 0 ( actual ) else ( ) - my-bulk-qh error? if ( ) - 0 ( actual ) \ USB error - else ( ) - my-bulk-qtd dup my-#qtds get-actual ( qtd actual ) - over >qtd-buf l@ rot >qtd-pbuf l@ 2 pick dma-sync ( actual ) - then ( actual ) + my-bulk-qtd dup my-#qtds get-actual ( qtd actual ) + over >qtd-buf l@ rot >qtd-pbuf l@ 2 pick dma-pull ( actual ) then ( actual ) usb-error ( actual usberr ) @@ -486,9 +485,7 @@ 0 instance value bulk-out-busy? : done-bulk-out ( -- error? ) \ Process results - my-bulk-qh done? 0= if my-bulk-qh error? drop then - - usb-error ( usberr ) + my-bulk-qh done-error? ( usberr ) my-bulk-qtd map-out-bptrs ( usberr ) my-bulk-qh fixup-bulk-out-data ( usberr ) my-bulk-qh remove-qh ( usberr ) @@ -507,11 +504,7 @@ my-bulk-qh >hcqh-overlay >hcqtd-token dup le-l@ TD_STAT_PING or swap le-l! \ OUT qTDs - TD_PID_OUT my-bulk-qtd fill-bulk-io-qtds ( ) - - \ Start bulk out transaction - my-bulk-qh pt-bulk fill-qh ( ) - my-bulk-qh insert-qh ( ) + TD_PID_OUT start-bulk-transaction ( ) true to bulk-out-busy? ( ) 0 ( usberr ) ; Modified: dev/usb2/hcd/ehci/control.fth ============================================================================== --- dev/usb2/hcd/ehci/control.fth Tue Oct 11 23:59:39 2011 (r2590) +++ dev/usb2/hcd/ehci/control.fth Wed Oct 12 00:28:40 2011 (r2591) @@ -135,19 +135,16 @@ \ Start control transaction my-qh pt-ctrl fill-qh + my-qh interrupt-on-last-td my-qh insert-qh \ Process results - my-qh done? if - 0 ( actual ) \ System error, timeout - else - my-qh error? if - 0 ( actual ) \ USB error - else - my-qtd >qtd-next l@ dup my-#qtds get-actual ( qtd actual ) - over >qtd-buf l@ rot >qtd-pbuf l@ 2 pick dma-sync ( actual ) - then - then + my-qh done-error? if ( ) + 0 ( actual ) \ System error, timeout, or USB error + else ( ) + my-qtd >qtd-next l@ dup my-#qtds get-actual ( qtd actual ) + over >qtd-buf l@ rot >qtd-pbuf l@ 2 pick dma-pull ( actual ) + then ( actual ) my-qh dup remove-qh free-qh ( actual ) usb-error ( actual usberr ) @@ -156,13 +153,12 @@ : run-control ( -- usberr ) \ Start control transaction my-qh pt-ctrl fill-qh + my-qh interrupt-on-last-td my-qh insert-qh \ Process results - my-qh done? 0= if my-qh error? drop then - - my-qh dup remove-qh free-qh - usb-error + my-qh done-error? ( usberr ) + my-qh dup remove-qh free-qh ( usberr ) ; : (control-set) ( sbuf sphy slen buf phy len -- usberr ) Modified: dev/usb2/hcd/ehci/ehci.fth ============================================================================== --- dev/usb2/hcd/ehci/ehci.fth Tue Oct 11 23:59:39 2011 (r2590) +++ dev/usb2/hcd/ehci/ehci.fth Wed Oct 12 00:28:40 2011 (r2591) @@ -69,6 +69,16 @@ usbsts@ dup usbsts! \ Clear interrupts and errors h# 10 and if " Host system error" USB_ERR_HCHALTED set-usb-error then ; +: hc-interrupt? ( -- interrupt? ) + usbsts@ h# 13 and dup if ( status ) + dup usbsts! ( status ) \ Clear interrupts, frame rollover, errors + dup h# 10 and if ( status ) + " USB host controller halted" USB_ERR_HCHALTED set-usb-error + then ( status ) + then ( status ) + 0<> ( interrupt? ) +; + : get-hc-status ( -- status ) usbsts@ dup usbsts! \ Clear interrupts and errors dup h# 10 and if " Host system error" USB_ERR_HCHALTED set-usb-error then Modified: dev/usb2/hcd/ehci/intr.fth ============================================================================== --- dev/usb2/hcd/ehci/intr.fth Tue Oct 11 23:59:39 2011 (r2590) +++ dev/usb2/hcd/ehci/intr.fth Wed Oct 12 00:28:40 2011 (r2591) @@ -36,6 +36,8 @@ 2 pick over d# 16 << or ( dir qtd /bptr token ) TD_C_ERR3 or TD_STAT_ACTIVE or ( dir qtd /bptr token' ) intr-in-data@ toggle-intr-in-data or ( dir qtd /bptr token' ) +\ See the comment about Interrupt-On-Completion elsewhere in this file +\ i my-#qtds 1- = if TD_IOC or then ( dir qtd /bptr token' ) 2 pick >hcqtd-token le-l! ( dir qtd /bptr ) my-buf++ ( dir qtd ) dup fixup-last-qtd ( dir qtd ) @@ -66,18 +68,28 @@ : intr-in? ( -- actual usberr ) intr-in-qh 0= if 0 USB_ERR_INV_OP exit then ( ) clear-usb-error ( ) - intr-in-qh qh-done? if ( ) - intr-in-qh error? if ( ) - 0 ( actual ) + \ Ironically, we can't use Interrupt-On-Completion for interrupt pipes, + \ because when a bulk or control transaction is performed with IOC, the + \ call to hc-interrupt? from that transaction can prevent this routine + \ from seeing its IOC completion status. The "right" solution might be to + \ scan all queue heads on an IOC "interrupt", then set a flag in the private + \ area of the queue head. It's unclear that that is better for interrupt + \ pipes than just polling, since the polling typically happens on a timer + \ and thus is relatively infrequent. + \ If we were to use Interrupt-On-Completion for interrupt pipes, we would + \ use "intr-qh-done?" instead of "qh-done?" here. + intr-in-qh qh-done? 0= if ( ) + 0 usb-error ( actual usberr ) \ Timeout or system error + else ( ) + intr-in-qh qh-error? if ( ) + 0 ( actual ) \ USB error else ( ) - intr-in-qh sync-qhqtds ( ) + intr-in-qh pull-qhqtds ( ) intr-in-qtd intr-in-qh >qh-#qtds l@ get-actual ( actual ) - intr-in-qtd >qtd-buf intr-in-qtd >qtd-pbuf l@ 2 pick dma-sync ( actual ) + intr-in-qtd >qtd-buf intr-in-qtd >qtd-pbuf l@ 2 pick dma-pull ( actual ) then ( actual ) usb-error ( actual usberr ) intr-in-qh fixup-intr-in-data ( actual usberr ) - else ( ) - 0 usb-error ( actual usberr ) then ( actual usberr ) ; @@ -87,6 +99,8 @@ dup >hcqtd-bptr0 dup le-l@ h# ffff.f000 and swap le-l! dup >qtd-/buf l@ d# 16 << TD_STAT_ACTIVE or TD_C_ERR3 or TD_PID_IN or +\ See the comment about Interrupt-On-Completion elsewhere in this file +\ TD_IOC or intr-in-data@ or over >hcqtd-token le-l! >qtd-next l@ @@ -104,7 +118,7 @@ intr-in-timeout intr-in-qh >qh-timeout l! intr-in-qh >hcqh-endp-char dup le-l@ QH_TD_TOGGLE invert and swap le-l! intr-in-qtd >qtd-phys l@ intr-in-qh >hcqh-overlay >hcqtd-next le-l! - intr-in-qh sync-qhqtds + intr-in-qh push-qhqtds ; : end-intr-in ( -- ) Modified: dev/usb2/hcd/ehci/iso.fth ============================================================================== --- dev/usb2/hcd/ehci/iso.fth Tue Oct 11 23:59:39 2011 (r2590) +++ dev/usb2/hcd/ehci/iso.fth Wed Oct 12 00:28:40 2011 (r2591) @@ -174,7 +174,7 @@ else \ Link and replace the entry itd-struct >itd-link le-l! ( itd-phys ) ( R: idx ) then - itd-struct over /itd-sync dma-sync ( itd-phys ) ( R: idx ) + itd-struct over /itd-sync dma-push ( itd-phys ) ( R: idx ) TYP_ITD or r> framelist! ( ) ; @@ -246,7 +246,7 @@ clear-usb-error iso-in-cur-itd set-itd-struct ( adr len ) itd-struct >itd-frame @ wait-frame-safe ( adr len ) - itd-struct dup >itd-phys @ /itd-sync dma-sync + itd-struct dup >itd-phys @ /itd-sync dma-pull iso-in-uframe-cnt 7 and ( adr len uframe# ) 4 * itd-struct >itd-xstat + le-l@ ( adr len xstat ) dup ITD_ACTIVE and if 3drop false exit then \ Not done yet @@ -295,14 +295,14 @@ [ifndef] notdef \ Debug aids -: ldump ( adr len -- ) " ldump" evaluate ; +: sys-ldump ( adr len -- ) " ldump" evaluate ; : .itd ( -- ) - iso-in-itd iso-in-#itd /itd-struct * ldump + iso-in-itd iso-in-#itd /itd-struct * sys-ldump ; : .xstat ( -- ) iso-in-#itd 0 do i set-itd-struct - itd-struct >itd-xstat /itd-xstat ldump + itd-struct >itd-xstat /itd-xstat sys-ldump loop ; \ Debug aids for video streaming only (and only if ITD_ACTIVE are all 0s) Modified: dev/usb2/hcd/ehci/qhtd.fth ============================================================================== --- dev/usb2/hcd/ehci/qhtd.fth Tue Oct 11 23:59:39 2011 (r2590) +++ dev/usb2/hcd/ehci/qhtd.fth Wed Oct 12 00:28:40 2011 (r2591) @@ -189,10 +189,14 @@ \ QH and its list of TDs are allocated as needed. \ --------------------------------------------------------------------------- -: sync-qh ( qh -- ) dup >qh-phys l@ /hcqh dma-sync ; -: sync-qtd ( qtd -- ) dup >qtd-phys l@ /hcqtd dma-sync ; -: sync-qtds ( qtd -- ) dup >qtd-phys l@ over >qtd-size l@ dma-sync ; -: sync-qhqtds ( qh -- ) dup >qh-phys l@ over >qh-size l@ dma-sync ; +: push-qh ( qh -- ) dup >qh-phys l@ /hcqh dma-push ; +: pull-qh ( qh -- ) dup >qh-phys l@ /hcqh dma-pull ; +: push-qtd ( qtd -- ) dup >qtd-phys l@ /hcqtd dma-push ; +: pull-qtd ( qtd -- ) dup >qtd-phys l@ /hcqtd dma-pull ; +: push-qtds ( qtd -- ) dup >qtd-phys l@ over >qtd-size l@ dma-push ; +: pull-qtds ( qtd -- ) dup >qtd-phys l@ over >qtd-size l@ dma-pull ; +: push-qhqtds ( qh -- ) dup >qh-phys l@ over >qh-size l@ dma-push ; +: pull-qhqtds ( qh -- ) dup >qh-phys l@ over >qh-size l@ dma-pull ; : map-out-bptrs ( qtd -- ) dup >qtd-buf l@ over >qtd-pbuf l@ rot >qtd-/buf-all l@ hcd-map-out @@ -214,27 +218,6 @@ TERMINATE swap >hcqtd-next-alt le-l! ( ) ; -: alloc-qtds ( #qtds -- qtd ) - dup >r /qtd * dup >r ( len ) ( R: #qtds len ) - aligned32-alloc-map-in ( u v p ) ( R: #qtds len ) - swap ( u p v ) ( R: #qtds len ) - dup r@ erase ( u p v ) ( R: #qtds len ) - - \ Record QTD size for later freeing - rot over >qtd-unaligned l! ( p v ) ( R: #qtds len ) - r> over >qtd-size l! ( p v ) ( R: #qtds ) - - dup rot r> link-qtds ( qtd.v ) -; - -: free-qtds ( qtd -- ) - >r ( R: qtd ) - r@ >qtd-unaligned l@ ( u ) ( R: qtd ) - r@ dup >qtd-phys l@ ( u v p ) ( R: qtd ) - r> >qtd-size l@ ( u v p size ) - aligned32-free-map-out ( ) -; - : link-qhqtd ( qtd.p qh -- ) >hcqh-overlay tuck ( qh.overlay qtd.p qh.overlay ) >hcqtd-next le-l! ( qh.overlay ) @@ -261,14 +244,6 @@ >qh-unaligned l! ( ) ; -: alloc-qh ( -- qh ) - /qh aligned32-alloc-map-in ( u v p ) - over /qh erase ( u v p ) - over >r ( u v p r: v ) - /qh 0 init-qh ( r: v ) - TERMINATE r@ link-qhqtd ( r: v ) - r> ( qh.v ) -; : free-qh ( qh -- ) >r ( R: qh ) r@ >qh-unaligned l@ ( qh.u ) ( R: qh ) @@ -375,9 +350,22 @@ : link-to-qh-ptr ( qh -- ) dup qh-ptr >qh-next l! ( qh ) dup >qh-phys l@ TYP_QH or qh-ptr >hcqh-next le-l! ( qh ) - sync-qhqtds ( ) + push-qhqtds ( ) ; +: interrupt-on-last-td ( qh -- ) + /qh + ( qtd ) + begin ( qtd ) + dup >hcqtd-next le-l@ TERMINATE <> ( qtd flag ) + while ( qtd ) + >qtd-next l@ ( qtd' ) + repeat ( qtd ) + >hcqtd-token ( 'token ) + dup le-l@ TD_IOC or swap le-l! ( ) +; + : insert-qh ( qh -- ) + hc-interrupt? drop \ Clear any pending transfer-complete interrupts + 0 >hcqh-cur-pqtd off qh-ptr if ( qh ) \ If there is another qh, link the new qh to the existing qh head. qh-ptr over >qh-prev l! ( qh ) @@ -386,7 +374,7 @@ link-to-qh-ptr ( ) - qh-ptr sync-qh ( ) + qh-ptr push-qh ( ) else ( ) \ If there is no other qh, make it the head, link it to itself, \ and start the asynch schedule. @@ -431,7 +419,7 @@ dup >qh-prev l@ ?dup if ( qh prev.qh ) over >hcqh-next le-l@ over >hcqh-next le-l! over >qh-next l@ swap >qh-next l! - dup sync-qh + dup push-qh dup >qh-next l@ qh-ptr <> if dup >qh-prev l@ swap >qh-next l@ >qh-prev l! else @@ -442,7 +430,7 @@ qh-ptr >hcqh-endp-char dup le-l@ QH_HEAD or swap le-l! ( qh ) fix-wraparound-qh ( ) 0 qh-ptr >qh-prev l! ( ) - qh-ptr sync-qh + qh-ptr push-qh then ring-doorbell then @@ -540,40 +528,61 @@ ; : qtd-done? ( qtd -- done? ) - >hcqtd-token le-l@ ( token ) - - dup TD_STAT_HALTED and if ( token ) - drop true exit - then ( token ) - TD_STAT_ACTIVE and 0= ( done? ) + >hcqtd-token le-l@ TD_STAT_ACTIVE and 0= ; +[ifdef] notdef +\ This version looks for the ending condition in the transfer overlay : qh-done? ( qh -- done? ) - process-hc-status ( qh ) - dup sync-qh ( qh ) - >hcqh-cur-pqtd le-l@ dup if qtd-done? then ( done? ) -; - -true value delay? -: poll-delay ( -- ) d# 300 " us" evaluate ; -: done? ( qh -- usberr ) - delay? if poll-delay then - begin dup qh-done? 0= while ( qh ) - 1 ms - dup >qh-timeout ( qh timeout-adr ) - dup l@ 1- ( qh timeout-adr timeout' ) - dup rot l! ( qh timeout' ) - 0= if ( qh ) -delay? 0= if cr 7 emit 7 emit ." TIMEOUT" cr debug-me then - " Timeout" USB_ERR_TIMEOUT set-usb-error ( qh ) - drop ( ) - usb-error ( usberr ) - exit - then - repeat ( qh ) - drop ( ) + hc-interrupt? 0= if drop false exit then ( qh ) + dup pull-qh ( qh ) + dup >hcqh-cur-pqtd l@ 0= if ( qh ) + drop false exit ( -- done? ) + then ( qh ) + >hcqh-overlay ( qtd ) + dup >hcqtd-token le-l@ ( qtd token ) + dup TD_STAT_ACTIVE and if ( qtd token ) + 2drop false ( done? ) + else ( qtd token ) + TD_STAT_HALTED and if ( qtd ) + drop true ( done? ) + else ( qtd ) + >hcqtd-next le-l@ ( next ) + TERMINATE = ( done? ) + then ( done? ) + then ( done? ) +; +[then] - usb-error ( usberr ) +\ This version traverses the list of QTDs, looking for an ending condition +: qh-done? ( qh -- done? ) + dup pull-qh ( qh ) +\ d# 30 us ( qh ) + /qh + ( qtd ) + begin ( qtd ) + dup >hcqtd-token le-l@ ( qtd token ) + dup TD_STAT_ACTIVE and 0= ( qtd token qtd-complete? ) + while ( qtd token ) + \ This QTD is complete - it is the end of the list if either + \ the next field is TERMINATE or if the HALTED bit is set + TD_STAT_HALTED and if ( qtd ) + drop true exit ( -- done? ) + then ( qtd ) + dup >hcqtd-next le-l@ TERMINATE = if ( qtd ) + drop true exit ( -- done? ) + then ( qtd ) + /qtd + ( qtd' ) + repeat ( qtd token ) + \ If we get here, we hit a QTD that is still active + 2drop false ( done? ) +; + +: intr-qh-done? ( qh -- done? ) + hc-interrupt? if ( qh ) + qh-done? ( done? ) + else ( qh ) + drop false ( done? ) + then ( done? ) ; : qtd-error? ( qtd qh -- usberr ) @@ -583,11 +592,27 @@ usb-error ; -: error? ( qh -- usberr ) dup >hcqh-overlay swap qtd-error? ; +: qh-error? ( qh -- usberr ) dup >hcqh-overlay swap qtd-error? ; + +: done-error? ( qh -- usberr ) + dup >qh-timeout l@ get-msecs + ( qh timeout ) + begin over intr-qh-done? 0= while ( qh timeout ) + usb-error ?dup if ." USB ERROR " . cr then + dup get-msecs - 0< if ( qh timeout ) + " Timeout" USB_ERR_TIMEOUT set-usb-error ( qh timeout ) + 2drop ( ) + usb-error ( usberr ) + exit + then ( qh timeout ) + repeat ( qh timeout ) + drop ( qh ) + + qh-error? ( usberr ) +; : get-actual ( qtd #qtd -- actual ) 0 -rot 0 ?do ( actual qtd ) - dup sync-qtd ( actual qtd ) + dup pull-qtd ( actual qtd ) dup >hcqtd-token le-l@ dup TD_STAT_ACTIVE and 0= if over >qtd-/buf l@ ( actual qtd token len ) swap d# 16 >> h# 7fff and - ( actual qtd len' ) @@ -601,7 +626,7 @@ : qtd-get-actual ( qtd -- actual ) 0 swap begin ( actual qtd ) - dup sync-qtd ( actual qtd ) + dup pull-qtd ( actual qtd ) dup >hcqtd-token le-l@ dup TD_STAT_ACTIVE and 0= if over >qtd-/buf l@ ( actual qtd token len ) swap d# 16 >> h# 7fff and - ( actual qtd len' ) Modified: dev/usb2/hcd/ohci/bulk.fth ============================================================================== --- dev/usb2/hcd/ohci/bulk.fth Tue Oct 11 23:59:39 2011 (r2590) +++ dev/usb2/hcd/ohci/bulk.fth Wed Oct 12 00:28:40 2011 (r2591) @@ -79,7 +79,7 @@ ( ed ) dup >hced-tdhead le-l@ ( dir ed head ) rot ED_DIR_IN = if bulk-in-data@ else bulk-out-data@ then or over >hced-tdhead le-l! ( ed ) - ( ed ) sync-edtds + ( ed ) push-edtds ; : insert-my-bulk ( ed dir -- ) over fill-bulk-ed insert-bulk ; : insert-my-bulk-in ( ed -- ) ED_DIR_IN insert-my-bulk ; @@ -108,12 +108,12 @@ : bulk-in-ready? ( -- false | error true | buf actual 0 true ) clear-usb-error ( ) process-hc-status ( ) - bulk-in-ed dup sync-edtds ( ed ) + bulk-in-ed dup pull-edtds ( ed ) ed-done? if ( ) bulk-in-td error? ?dup 0= if ( ) bulk-in-td >td-cbp l@ ( buf ) bulk-in-td get-actual ( buf actual ) - 2dup bulk-in-td >td-pcbp l@ swap dma-sync ( buf actual ) + 2dup bulk-in-td >td-pcbp l@ swap dma-pull ( buf actual ) 0 ( buf actual 0 ) then ( error | buf actual 0 ) bulk-in-ed fixup-bulk-in-data ( error | buf actual 0 ) @@ -126,13 +126,13 @@ bulk-in-ed 0= if 0 USB_ERR_INV_OP exit then clear-usb-error ( ) process-hc-status ( ) - bulk-in-ed dup sync-edtds ( ed ) + bulk-in-ed dup pull-edtds ( ed ) ed-done? if ( ) bulk-in-td error? if 0 ( actual ) else bulk-in-td dup get-actual ( td actual ) - over >td-cbp l@ rot >td-pcbp l@ 2 pick dma-sync ( actual ) + over >td-cbp l@ rot >td-pcbp l@ 2 pick dma-pull ( actual ) then usb-error ( actual usberr ) bulk-in-ed fixup-bulk-in-data ( actual usberr ) @@ -165,7 +165,7 @@ \ Setup ED again bulk-in-td >td-phys l@ bulk-in-data@ or bulk-in-ed >hced-tdhead le-l! - bulk-in-ed dup sync-edtds + bulk-in-ed dup push-edtds ed-unset-skip enable-bulk ; @@ -198,7 +198,7 @@ 0 ( actual ) \ USB error else my-td dup get-actual ( td actual ) - over >td-cbp l@ rot >td-pcbp l@ 2 pick dma-sync ( actual ) + over >td-cbp l@ rot >td-pcbp l@ 2 pick dma-pull ( actual ) then then Modified: dev/usb2/hcd/ohci/control.fth ============================================================================== --- dev/usb2/hcd/ohci/control.fth Tue Oct 11 23:59:39 2011 (r2590) +++ dev/usb2/hcd/ohci/control.fth Wed Oct 12 00:28:40 2011 (r2591) @@ -105,7 +105,7 @@ : insert-my-control ( -- ) my-ed dup fill-control-ed - dup sync-edtds + dup push-edtds insert-control ; @@ -146,7 +146,7 @@ 0 ( actual ) \ USB error else my-td >td-next l@ dup get-actual ( td actual ) - over >td-cbp l@ rot >td-pcbp l@ 2 pick dma-sync ( actual ) + over >td-cbp l@ rot >td-pcbp l@ 2 pick dma-pull ( actual ) then then Modified: dev/usb2/hcd/ohci/edtd.fth ============================================================================== --- dev/usb2/hcd/ohci/edtd.fth Tue Oct 11 23:59:39 2011 (r2590) +++ dev/usb2/hcd/ohci/edtd.fth Wed Oct 12 00:28:40 2011 (r2591) @@ -45,7 +45,7 @@ \ Initialize hcca hcca /hcca erase - hcca hcca-phys /hcca dma-sync + hcca hcca-phys /hcca dma-push ; \ --------------------------------------------------------------------------- @@ -138,7 +138,7 @@ dup >hced-control dup le-l@ ED_SKIP_MASK invert and r> or swap le-l! - dup >ed-phys l@ /hced dma-sync + dup >ed-phys l@ /hced dma-push ; : ed-set-skip ( ed -- ) ED_SKIP_ON (set-skip) ; : ed-unset-skip ( ed -- ) ED_SKIP_OFF (set-skip) ; @@ -269,10 +269,15 @@ r> >ed-size l@ ( ed.u,v,p size ) aligned32-free-map-out ( ) ; -: sync-edtds ( ed -- ) +: push-edtds ( ed -- ) dup >ed-phys l@ ( ed.v,p ) over >ed-size l@ ( ed.v,p len ) - dma-sync ( ) + dma-push ( ) +; +: pull-edtds ( ed -- ) + dup >ed-phys l@ ( ed.v,p ) + over >ed-size l@ ( ed.v,p len ) + dma-pull ( ) ; : map-out-cbp ( td -- ) dup >td-cbp l@ over >td-pcbp l@ rot >td-/cbp-all l@ hcd-map-out @@ -429,7 +434,7 @@ : done? ( ed -- error? ) begin process-hc-status - dup sync-edtds + dup pull-edtds dup ed-done? ?dup 0= if 1 ms timeout 1- dup to timeout 0= Modified: dev/usb2/hcd/ohci/intr.fth ============================================================================== --- dev/usb2/hcd/ohci/intr.fth Tue Oct 11 23:59:39 2011 (r2590) +++ dev/usb2/hcd/ohci/intr.fth Wed Oct 12 00:28:40 2011 (r2591) @@ -52,7 +52,7 @@ ( ed ) dup >hced-tdhead le-l@ ( ed head ) intr-in-data@ or ( ed head' ) over >hced-tdhead le-l! ( ed ) - ( ed ) sync-edtds ( ) + ( ed ) push-edtds ( ) ; : remove-my-intr ( ed -- ) dup remove-intr free-edtds ; @@ -80,13 +80,13 @@ intr-in-ed 0= if 0 USB_ERR_INV_OP exit then clear-usb-error ( ) process-hc-status ( ) - intr-in-ed dup sync-edtds ( ed ) + intr-in-ed dup pull-edtds ( ed ) ed-done? if ( ) intr-in-td error? if 0 ( actual ) else ( ) intr-in-td dup get-actual ( td actual ) - over >td-cbp l@ rot >td-pcbp l@ 2 pick dma-sync ( actual ) + over >td-cbp l@ rot >td-pcbp l@ 2 pick dma-pull ( actual ) then usb-error ( actual usberr ) intr-in-ed fixup-intr-in-data ( actual usberr ) @@ -118,7 +118,7 @@ \ Setup ED again intr-in-td >td-phys l@ intr-in-data@ or intr-in-ed >hced-tdhead le-l! - intr-in-ed dup sync-edtds + intr-in-ed dup push-edtds ed-unset-skip ; Modified: dev/usb2/hcd/uhci/bulk.fth ============================================================================== --- dev/usb2/hcd/uhci/bulk.fth Tue Oct 11 23:59:39 2011 (r2590) +++ dev/usb2/hcd/uhci/bulk.fth Wed Oct 12 00:28:40 2011 (r2591) @@ -113,12 +113,12 @@ bulk-in-qh 0= if 0 USB_ERR_INV_OP exit then clear-usb-error process-hc-status - bulk-in-qh dup sync-qhtds ( bulk-in-qh ) + bulk-in-qh dup pull-qhtds ( bulk-in-qh ) qh-done? if ( ) bulk-in-actual ( actual usberr ) else ( ) bulk-in-qh dup >hcqh-elem le-l@ ( qh elem ) - 1 ms over sync-qhtds ( qh elem ) + 1 ms over pull-qhtds ( qh elem ) swap >hcqh-elem le-l@ = if \ No movement in the past ms bulk-in-actual ( actual usberr ) bulk-in-td bulk-in-qh >qh-#tds l@ fixup-bulk-in-data @@ -127,7 +127,7 @@ then then over if - bulk-in-td dup >td-buf l@ swap >td-pbuf l@ 2 pick dma-sync + bulk-in-td dup >td-buf l@ swap >td-pbuf l@ 2 pick dma-pull then ; @@ -155,7 +155,7 @@ \ Setup QH again bulk-in-td >td-phys l@ bulk-in-qh >hcqh-elem le-l! - bulk-in-qh sync-qhtds + bulk-in-qh push-qhtds ; : end-bulk-in ( -- ) @@ -188,7 +188,7 @@ 0 ( actual ) \ USB error else my-td dup my-#tds get-actual ( td actual ) - over >td-buf l@ rot >td-pbuf l@ 2 pick dma-sync ( actual ) + over >td-buf l@ rot >td-pbuf l@ 2 pick dma-pull ( actual ) then then Modified: dev/usb2/hcd/uhci/control.fth ============================================================================== --- dev/usb2/hcd/uhci/control.fth Tue Oct 11 23:59:39 2011 (r2590) +++ dev/usb2/hcd/uhci/control.fth Wed Oct 12 00:28:40 2011 (r2591) @@ -125,7 +125,7 @@ 0 ( actual ) \ USB error else my-td >td-next l@ dup my-#tds get-actual ( td actual ) - over >td-buf l@ rot >td-pbuf l@ 2 pick dma-sync ( actual ) + over >td-buf l@ rot >td-pbuf l@ 2 pick dma-pull ( actual ) then then Modified: dev/usb2/hcd/uhci/intr.fth ============================================================================== --- dev/usb2/hcd/uhci/intr.fth Tue Oct 11 23:59:39 2011 (r2590) +++ dev/usb2/hcd/uhci/intr.fth Wed Oct 12 00:28:40 2011 (r2591) @@ -76,11 +76,11 @@ intr-in-qh 0= if 0 USB_ERR_INV_OP exit then clear-usb-error process-hc-status - intr-in-qh dup sync-qhtds + intr-in-qh dup pull-qhtds qh-done? if intr-in-td intr-in-qh >qh-#tds l@ get-actual ( actual ) usb-error ( actual usberr ) - intr-in-td dup >td-buf l@ swap >td-pbuf l@ 2 pick dma-sync + intr-in-td dup >td-buf l@ swap >td-pbuf l@ 2 pick dma-pull intr-in-td intr-in-qh >qh-#tds l@ fixup-intr-in-data else 0 usb-error ( actual usberr ) @@ -109,7 +109,7 @@ \ Setup QH again intr-in-td >td-phys l@ intr-in-qh >hcqh-elem le-l! - intr-in-qh sync-qhtds + intr-in-qh push-qhtds ; : end-intr-in ( -- ) Modified: dev/usb2/hcd/uhci/qhtd.fth ============================================================================== --- dev/usb2/hcd/uhci/qhtd.fth Tue Oct 11 23:59:39 2011 (r2590) +++ dev/usb2/hcd/uhci/qhtd.fth Wed Oct 12 00:28:40 2011 (r2591) @@ -230,8 +230,12 @@ aligned16-free-map-out ; -: sync-qhtds ( qh -- ) - dup >qh-phys l@ over >qh-size l@ dma-sync +: push-qhtds ( qh -- ) + dup >qh-phys l@ over >qh-size l@ dma-push +; + +: pull-qhtds ( qh -- ) + dup >qh-phys l@ over >qh-size l@ dma-pull ; : map-out-buf ( td -- ) @@ -299,7 +303,7 @@ process-hc-status usb-error if true else - dup sync-qhtds + dup pull-qhtds dup qh-done? ?dup 0= if 1 ms timeout 1- dup to timeout 0= @@ -310,7 +314,7 @@ ( qh ) dup qh-done? 0= if " Timeout" USB_ERR_TIMEOUT set-usb-error TERMINATE over >hcqh-elem le-l! \ Terminate QH - sync-qhtds + push-qhtds 1 ms else drop

1 0

[commit] r2590 - dev/usb2/device/keyboard
by repository service 11 Oct '11

11 Oct '11

Author: wmb Date: Tue Oct 11 23:59:39 2011 New Revision: 2590 URL: http://tracker.coreboot.org/trac/openfirmware/changeset/2590 Log: USB keyboard driver - flash the keyboard LEDs on open as a clue that the keyboard has been attached. Modified: dev/usb2/device/keyboard/kbd.fth Modified: dev/usb2/device/keyboard/kbd.fth ============================================================================== --- dev/usb2/device/keyboard/kbd.fth Tue Oct 11 22:19:10 2011 (r2589) +++ dev/usb2/device/keyboard/kbd.fth Tue Oct 11 23:59:39 2011 (r2590) @@ -339,6 +339,8 @@ \ Some USB keyboards don't implement set-idle properly, and it's not critical, \ so we suppress the message to avoid confusing the user idle-rate set-idle drop \ if ." Failed to set USB keyboard idle" cr then + 7 set-leds \ Flash the LEDs to indicate that OFW has attached the keyboard + d# 200 ms 0 set-leds false ;

1 0

[commit] r2589 - cpu/x86/pc/olpc
by repository service 11 Oct '11

11 Oct '11

Author: wmb Date: Tue Oct 11 22:19:10 2011 New Revision: 2589 URL: http://tracker.coreboot.org/trac/openfirmware/changeset/2589 Log: OLPC - add some guards to load-from-list. Modified: cpu/x86/pc/olpc/security.fth Modified: cpu/x86/pc/olpc/security.fth ============================================================================== --- cpu/x86/pc/olpc/security.fth Tue Oct 11 20:17:44 2011 (r2588) +++ cpu/x86/pc/olpc/security.fth Tue Oct 11 22:19:10 2011 (r2589) @@ -1230,11 +1230,14 @@ ec-indexed-io-off ( list$ ) 2drop ( ) ['] secure-load-ramdisk to load-ramdisk - " init-program" $find if -\ set-cmdline - execute - sound-end - go + " init-program" $find if ( xt ) +\ set-cmdline ( xt ) + catch ?dup if ( error ) + .error ( ) + else ( ) + sound-end ( ) + ['] go guarded ( ) + then then show-x security-failure

1 0

← Newer
1
...
4
5
6
7
8
9
10
11
Older →

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

openfirmware October 2011